Furthermore, allow OpenSSL configured with no-dh. Additionally, this change
allows auto-negotiation of the elliptic curve/group for servers, not only with
OpenSSL 1.0.2 but also with OpenSSL 1.1.0 and newer. This enables X25519
(since OpenSSL 1.1.0) and X448 (since OpenSSL 1.1.1) as a side-effect.
ASTERISK-27910
Change-Id: I5b0dd47c5194ee17f830f869d629d7ef212cf537
Certain race conditions between changing bridge types and DTMF can
cause the current FLAG_NEED_MARKER_BIT to send the marker bit before
the actual first packet of native bridging.
This logic keeps track of the ssrc the bridge is currently sending
and will correctly ensure the marker bit is set if SSRC as changed
from the previous sent packet.
ASTERISK-27845
Change-Id: I01858bd0235f1e5e629e20de71b422b16f55759b
When RTP was originally created it had the ability to place a single
extension in an RTP packet. In practice people wanted to potentially
put multiple extensions in one and so RFC 5285 (obsoleted by RFC
8285) came into existence. This allows RTP extensions to be negotiated
with a unique identifier to be used in the RTP packet, allowing
multiple extensions to be present in the packet.
This change extends the RTP engine API to add support for this. A
user of it can enable extensions and the API provides the ability to
retrieve the information (to construct SDP for example) and to provide
negotiated information (from SDP). The end result is that the RTP
engine can then query to see if the extension has been negotiated and
what unique identifier is to be used. It is then up to the RTP engine
implementation to construct the packet appropriately.
The first extension to use this support is abs-send-time which is
defined in the REMB draft[1] and is a second timestamp placed in an
RTP packet which is for when the packet has left the sending system.
It is used to more accurately determine the available bandwidth.
ASTERISK-27831
[1] https://tools.ietf.org/html/draft-alvestrand-rmcat-remb-03
Change-Id: I508deac557867b1e27fc7339be890c8018171588
When the local SSRC changes we need to update the SRTP information
so that the proper key is used. This is commonly done as a result
of bridging two channels together. Previously we only updated
the SRTP information if media had already flowed, but in practice
the channel driver may have already performed SRTP negotiation and
set up the previous SSRC. We now always do it on a local SSRC
change.
ASTERISK-27795
ASTERISK-27800
Change-Id: Ia7c8e74c28841388b5244ac0b8fd6c1dc6ee4c10
Adds the ability to receive and handle incoming NACK requests if
retransmissions are enabled. If retransmissions are enabled, a data
buffer is allocated that stores packets being sent. If a NACK request
is received, the packet requested for retransmission is sent if it is
still in the buffer. In the same request, if any of the following 16
packets are marked as not received, those will be sent as well if
available, as outlined in RFC4585.
Also changes RTCP RR and SR to use media source SSRC instead of packet
source SSRC when determining which instance to use for RTCP reports.
For more information, refer to the wiki page:
https://wiki.asterisk.org/wiki/display/AST/WebRTC+User+Experience+Improvements
ASTERISK-27806 #close
Change-Id: I7f7f124af3b9d5d2fd9cffc6ba8cb48a6fff06ec
This change allows chan_pjsip to be given an AST_FRAME_RTCP
containing REMB feedback and pass it to res_rtp_asterisk.
Once res_rtp_asterisk receives the frame a REMB RTCP feedback
packet is constructed with the appropriate contents and sent
to the remote endpoint.
ASTERISK-27776
Change-Id: Ic53f821c1560d8924907ad82c4d9c0bc322b38cd
The previous payload specific feedback handling was very single
minded in that it just assumed everything should trigger a video
update. This was changed but the handling of picture loss indication
was not added. The result was that video may not flow. This change
adds it explicitly in.
Change-Id: I1894be02e39ee10a0af841b5a1dca5f0ec7d60b6
This change extends the existing AST_FRAME_RTCP frame type to be
able to contain additional RTCP message types, such as feedback
messages. The payload type is contained in the subclass which allows
knowing what is in the frame itself.
The RTCP feedback message type is now handled and REMB[1] messages
are raised with their containing information.
This also fixes a bug where all feedback messages were triggering
video updates instead of just FIR and FUR.
Finally RTCP frames are now passed up through the Asterisk core to
what is handling the channel, mapped appropriately in the case of
bridging, and written to an outgoing stream. Since RTCP frames are
on a per-stream basis this is only done on multistream capable
channels.
[1] https://tools.ietf.org/html/draft-alvestrand-rmcat-remb-03
ASTERISK-27758
ASTERISK-26366
Change-Id: I680da0ad8d5059d5e9655d896fb9d92e9da8491e
Checking option_debug directly is incorrect as it ignores file/module
specific debug settings. This system-wide change replaces nearly all
direct checks for option_debug with the DEBUG_ATLEAST macro.
Change-Id: Ic342d4799a945dbc40ac085ac142681094a4ebf0
The pool cache gets in the way of finding use after free errors of memory
pool contents. Tools like valgrind and MALLOC_DEBUG don't know when a
pool is released because it gets put into the cache instead of being
freed.
* Added the "cache_pools" option to pjproject.conf. Disabling the option
helps track down pool content mismanagement when using valgrind or
MALLOC_DEBUG. The cache gets in the way of determining if the pool
contents are used after free and who freed it.
To disable the pool caching simply disable the cache_pools option in
pjproject.conf and restart Asterisk.
Sample pjproject.conf setting:
[startup]
cache_pools=no
* Made current users of the caching pool factory initialization and
destruction calls call common routines to create and destroy cached pools.
ASTERISK-27704
Change-Id: I64d5befbaeed2532f93aa027a51eb52347d2b828
If the ICE role is not set right away, we might have a role conflict
that stays undetected and ICE finishing with successful tests and no
candidate nominated. This was introduced by ASTERISK-27088.
To avoid this, we set the role as soon as before but only if the ICE
state permits it: still checking and not yet nominating candidates or
completed.
ASTERISK-27646
Change-Id: I5dbc69ad63cacbb067922850fbb113d479bd729c
When RTCP-MUX enabled. rtp->s is the same as rtcp->s, check this before
close the file descriptor. Close the FD twice will hangs the asterisk
under heavy load.
ASTERISK-27299 #close
Reported-by: Aaron An
Tested-by: AaronAn
Change-Id: I870a072d73fd207463ac116ef97100addbc0820a
We should not do flood detection on video RTP streams. Video RTP streams
are very bursty by nature. They send out a burst of packets to update the
video frame then wait for the next video frame update. Really only audio
streams can be checked for flooding. The others are either bursty or
don't have a set rate.
* Added code to selectively disable packet flood detection for video RTP
streams.
ASTERISK-27440
Change-Id: I78031491a6e75c2d4b1e9c2462dc498fe9880a70
When the RTCP code was transitioned over to Stasis a code change
was made to keep track of how many reports are present. This count
controlled where report blocks were placed in the RTCP report.
If a compound RTCP packet was received this logic would incorrectly
place a report block in the wrong location resulting in a write
to an invalid location.
This change removes this counting logic and always places the report
block at the first position. If in the future multiple reports are
supported the logic can be extended but for now keeping a count
serves no purpose.
ASTERISK-27382
ASTERISK-27429
Change-Id: Iad6c8a9985c4b608ef493e19c421211615485116
There are many places in the code base where we ignore the return value
of fcntl() when getting/setting file descriptior flags. This patch
introduces a convenience function that allows setting or clearing file
descriptor flags and will also log an error on failure for later
analysis.
Change-Id: I8b81901e1b1bd537ca632567cdb408931c6eded7
More complicated direct media reinvite negotiations can result in longer
delays before direct media flows. The strictrtp learning timeout time
was too short. One log showed that the first RTP packet came in just
after three seconds.
* Increase the strictrtp learning timeout time from 1.5 to 5 seconds.
ASTERISK-27453
Change-Id: Ic5e711164cbb91b4d1c1e40c83697755640f138c
Previously, Asterisk sent srflx only when configured exclusively for IPv4. Now,
srflx is gathered and sent via SDP, even when Asterisk is enabled for
Dual Stack (IPv4+IPv6) and an IPv4 interface is available/used.
ASTERISK-27437
Change-Id: Ie07d8e2bfa7b6fe06fcdc73d390a7a9a4d8c0bc1
Some clients do not send rtp packets every ptime ms. This can lead to
situations in which the rtp source learning algorithm will never learn
the address of the client. This has been discovered on a Mac mini with
a pjsip based softphone after updating to Sierra: as soon as USB
headsets are involved, the softphone will send the second packet 30ms
after the first, the third 30ms after the second and the fourth 1ms
after the third. So in the old implmentation the rtp source learning
algorithm was repeatedly reset on the fourth packet.
The patch changes the algorithm in a way that doesn't take the arrival
time between two consecutive packets into account but the time between
the first and the last packet of a learning sequence.
The patch also fixes a second problem: when a user was using a wrong
value for the probation setting there was a LOG_WARNING output stating
that the value had been set to the default value instead. However
the code for setting the value back to defaults was missing.
ASTERISK-27421 #close
Change-Id: If778fe07678a6fd2041eaca7cd78267d0ef4fc6c
This mimics the behavior of Chrome and Firefox and creates an ephemeral
X.509 certificate for each DTLS session.
Currently, the only supported key type is ECDSA because of its faster
generation time, but other key types can be added in the future as
necessary.
ASTERISK-27395
Change-Id: I5122e5f4b83c6320cc17407a187fcf491daf30b4
The bridge_p2p_rtp_write() has potential reentrancy problems.
* Accessing the bridged RTP members must be done with the instance1 lock
held. The DTMF and asymmetric codec checks must be split to be done with
the correct RTP instance struct locked. i.e., They must be done when
working on the appropriate side of the point to point bridge.
* Forcing the RTP mark bit was referencing the wrong side of the point to
point bridge. The set mark bit is used everywhere else to set the mark
bit when sending not receiving.
The patches for ASTERISK_26745 and ASTERISK_27158 did not take into
account that not everything carried by RTP uses a codec. The telephony
DTMF events are not exchanged with a codec. As a result when
RFC2833/RFC4733 sent digits you would crash if "core set debug 1" is
enabled, the DTMF digits would always get passed to the core even though
the local native RTP bridge is active, and the DTMF digits would go out
using the wrong SSRC id.
* Add protection for non-format payload types like DTMF when updating the
lastrxformat and lasttxformat. Also protect against non-format payload
types when checking for asymmetric codecs.
ASTERISK-27292
Change-Id: I6344ab7de21e26f84503c4d1fca1a41579364186
This could have been fixed by subtracting 1 from the final value of
'len' but the way the packet was being constructed was confusing so I
took the opportunity to (I think) make it more clear.
We were sending 1 extra byte at the end of the SDES RTCP packet which
caused Chrome to complain (in its debug log):
Too little data (1 byte) remaining in buffer to parse
RTCP header (4 bytes).
We now send the correct number of bytes.
Change-Id: I9dcf087cdaf97da0374ae0acb7d379746a71e81b
Assertions in the v15+ AST-2017-008 patches found that we were not
handling the case if the incoming SDP did not specify the required SSRC
attributes for bundled to work.
* Be strict on matching SSRC for bundled instances including the parent
instance. If the SSRC doesn't match then discard the packet. Bundled has
to tell us in the SDP signaling what SSRC to expect. Otherwise, we will
not know how to find the bundled instance structure.
Change-Id: I152830bbff71c662408909042068fada39e617f9
Validate RTCP packets before processing them.
* Validate that the received packet is of a minimum length and apply the
RFC3550 RTCP packet validation checks.
* Fixed potentially reading garbage beyond the received RTCP record data.
* Fixed rtp->themssrc only being set once when the remote could change
the SSRC. We would effectively stop handling the RTCP statistic records.
* Fixed rtp->themssrc to not treat a zero value as special by adding
rtp->themssrc_valid to indicate if rtp->themssrc is available.
ASTERISK-27274
Make strict RTP learning more flexible.
Direct media can cause strict RTP to attempt to learn a remote address
again before it has had a chance to learn the remote address the first
time. Because of the rapid relearn requests, strict RTP could latch onto
the first remote address and fail to latch onto the direct media remote
address. As a result, you have one way audio until the call is placed on
and off hold.
The new algorithm learns remote addresses for a set time (1.5 seconds)
before locking the remote address. In addition, we must see a configured
number of remote packets from the same address in a row before switching.
* Fixed strict RTP learning from always accepting the first new address
packet as the new stream.
* Fixed strict RTP to initialize the expected sequence number with the
last received sequence number instead of the last transmitted sequence
number.
* Fixed the predicted next sequence number calculation in
rtp_learning_rtp_seq_update() to handle overflow.
ASTERISK-27252
Change-Id: Ia2d3aa6e0f22906c25971e74f10027d96525f31c
This change moves the logic which learns a new source address
for RTP so it only occurs in the learning state. The learning
state is entered on initial allocation of RTP or if we are
told that the remote address for the media has changed. While
in the learning state if we continue to receive media from
the original source we restart the learning process. It is
only once we receive a sufficient number of RTP packets from
the new source that we will switch to it. Once this is done
the closed state is entered where all packets that do not
originate from the expected source are dropped.
The learning process has also been improved to take into
account the time between received packets so a flood of them
while in the learning state does not cause media to be switched.
Finally RTCP now drops packets which are not for the learned
SSRC if strict RTP is enabled.
ASTERISK-27013
Change-Id: I56a96e993700906355e79bc880ad9d4ad3ab129c
When SDP renegotiation occurs it is possible for an RTP
instance to be reused for a new stream, resulting in the remote
SSRC changing if it is part of a bundle group. This change
allows this and updates its mapping in the current bundle
group.
ASTERISK-27231
Change-Id: I6e3703974f236bc024c5dbe9bd43adae0c6fb490
Asterisk wasn't generating or forwarding RTCP packets when native
bridge was activated. Also the stats weren't available via
CHANNEL(qos). Now the RTCP stats are always calculated.
ASTERISK-27158 #close
Change-Id: I46fb8f61c95e836b9d2dda6054b0cf205c16037b
Introduce a new property to rtp-engine to make it aware of
the desire for assymetric codecs or not. If asymmetric codecs
is not allowed, the bridge will compare read/write formats
and shut down the p2p bridge if needed
ASTERISK-26745 #close
Change-Id: I0d9c83e5356df81661e58d40a8db565833501a6f
This change does a few things to improve packet loss and renegotiation:
1. On outgoing RTP streams we will now properly reflect out of order
packets and packet loss in the sequence number. This allows the
remote jitterbuffer to better reorder things.
2. Video updates can now be discarded for a period of time
after one has been sent to prevent flooding of clients.
3. For declined and removed streams we will now release any
media session resources associated with them. This was not
previously done and caused an issue where old state was being
used for a new stream.
4. RTP bundling was not actually removing bundled RTP instances
from the parent. This has been resolved by removing based on
the RTP instance itself and not the SSRC.
5. The code did not properly handle explicitly unbundling an
RTP instance from its parent. This now works as expected.
ASTERISK-27143
Change-Id: Ibd91362f0e4990b6129638e712bc8adf0899fd45
This change makes it so that if an RTCP packet is being sent
the RTP ICE component is used for sending if RTCP-MUX is in use.
ASTERISK-27133
Change-Id: I6200f611ede709602ee9b89501720c29545ed68b
BUNDLE is a specification used in WebRTC to allow multiple
streams to use the same underlying transport. This reduces
the number of ICE and DTLS negotiations that has to occur
to 1 normally.
This change implements this by adding support for it to
the RTP SDP module in PJSIP. BUNDLE can be turned on using
the "bundle" option and on an offer we will offer to
bundle streams together. On an answer we will accept any
bundle groups provided. Once accepted each stream is bundled
to another RTP instance for transport.
For the res_rtp_asterisk changes the ability to bundle
an RTP instance to another based on the SSRC received
from the remote side has been added. For outgoing traffic
if an RTP instance is bundled to another we will use the
other RTP instance for any transport related things. For
incoming traffic received from the transport instance we
look up the correct instance based on the SSRC and use it
for any non-transport related data.
ASTERISK-27118
Change-Id: I96c0920b9f9aca7382256484765a239017973c11
When a message is received on the TURN socket, the code processing the
message needs to call into the ICE/STUN session for further processing.
This code path locks the TURN group lock then the ICE/STUN group lock. In
another thread an ICE/STUN timer can fire off to send a keep alive message
over the TURN socket. In this code path, the ICE/STUN group lock is
obtained then the TURN group lock is obtained to send the packet. A
classic deadlock case if the group locks are not the same.
* Made TURN get created using the ICE/STUN session's group lock.
NOTE: I was originally concerned that the ICE/STUN session can get
recreated by ice_reset_session() for an event like RTCP multiplexing
causing a change during SDP negotiation. In this case the TURN group lock
would become different. However, TURN is also recreated as part of the
ICE/STUN recreation in ice_create() when all known ICE candidates are
added to the new ICE session. While the ICE/STUN and TURN sessions are
being recreated there is a period where the group locks could be
different.
ASTERISK-27023 #close
Patches:
res_rtp_asterisk-turn-deadlock-fix.patch (license #6502)
patch uploaded by Michael Walton (modified)
Change-Id: Ic870edb99ce4988a8c8eb6e678ca7f19da1432b9
There needed to be a way to notify handlers upstream that DTLS had been
established. This patch makes it so once DTLS has been estalished a source
change control frame is put into the read queue. Any handlers can then watch
for that frame and trigger off of it.
ASTERISK-27096 #close
Change-Id: I27ff344f5a8c691a1890dfe3254a4b1a49e7f4a0
When re-inviting to add more streams it is possible for
the role of existing ICE sessions to be changed to the
incorrect value. This results in subsequent refreshes
within the sessions getting a role conflict and the ICE
session breaking down. This change only sets the role to
be the new value if an ICE renegotiation is actually
going to happen, otherwise the existing role is preserved.
As well if we encounter a situation where a unidirectional
ICE negotiation happens and the other side does not send us
candidates we will not store any information for sending
traffic, even though we know where they are reachable. This
change fixes this by using the source of the ICE traffic
itself as the target if no candidates are known and we
receive some ICE traffic.
ASTERISK-27088
Change-Id: I71228181e358917fcefc3100fad21b2fc02a59a9
It looks like there was a copy/paste error in ast_rtp_change_source
where if there was a rtcp srtp instance, instead of updating its
ssrc we were updating the srtp instance ssrc twice.
ASTERISK-27022 #close
Reported-by: Michael Walton
Change-Id: Ic88f3aee7227b401c58745ac265ff92c19620095
In review 4843 (ASTERISK-24858), we added a hack that forced a smoother
creation when sending signed linear so that the byte order was adjusted
during transmission. This was needed because smoother flags were lost
during the new format work that was done in Asterisk 13.
Rather than rolling that same hack into res_rtp_multicast, re-introduce
smoother flags so that formats can dictate their own options.
Change-Id: I77b835fba0e539c6ce50014a984766f63cab2c16
When using rtcp mux if an rtcp payload came in it would still use the srtp
unprotect algorithm instead of the srtp unprotect rtcp method. Since rtcp
data was being passed to the rtp unprotect method this would result in an
error.
This patch ensures that the correct unprotect method is chosen by making
sure the passed in rtcp flag is appropriately set when rtcp mux is enabled
and an rtcp payload is received.
ASTERISK-26979 #close
Change-Id: Ic5409f9d1a267f1d4785fc5aed867daaecca6241
When a call gets put on hold RTP is temporarily stopped and Asterisk was
setting the remote RTCP address to NULL. Then when RTCP data was received
from the remote endpoint, Asterisk would be missing this information when
publishing the rtcp_message stasis event. Consequently, message subscribers
(in this case res_hep_rtcp) trying to parse the "from" field output the
following error:
"ast_sockaddr_split_hostport: Port missing in (null)"
This patch makes it so the remote RTCP address is no longer set to NULL when
stopping RTP. There was only one place that appeared to check if the remote
RTCP address was NULL as a way to tell if RTCP was running. This patch added
an additional check on the RTCP schedid for that case to make sure RTCP was
truly not running.
ASTERISK-26860 #close
Change-Id: I6be200fb20db647e48b5138ea4b81dfa7962974b
